SwarmUI Master Tutorial - Use Stable Diffusion 3.5 Large and FLUX model with Amazing Performance and more

In this public tutorial post I will share all the necessary stuff regarding how to use Stable Diffusion 3.5 Large, FLUX Dev, SD 1.5, SDXL, Stable Diffusion 3 and on your computer and also on the cloud (RunPod + Massed Compute).

=======

Patreon exclusive posts index to find our scripts easily, Patreon scripts updates history to see which updates arrived to which scripts and amazing Patreon special generative scripts list that you can use in any of your task.

Join discord to get help, chat, discuss and also tell me your discord username to get your special rank : SECourses Discord

Please also Star, Watch and Fork our Stable Diffusion & Generative AI  GitHub repository and join our Reddit subreddit and follow me on LinkedIn (my real profile)

=======

Amazing Ultra Important Tutorials with Chapters and Manually Written Subtitles / Captions

Updated Playlist : https://www.youtube.com/playlist?list=PL_pbwdIyffsku5Rfv2CJZSiFdGXtkkdiB

• FLUX Tools Outpainting, Inpainting (Fill), Redux, Depth & Canny Ultimate Tutorial Guide with SwarmUI

  • https://youtu.be/hewDdVJEqOQ

• Stable Diffusion 3.5 Large How To Use Tutorial With Best Configuration and Comparison With FLUX DEV

  • https://youtu.be/-zOKhoO9a5s

• FLUX Full Fine-Tuning / DreamBooth Tutorial That Shows A Lot Info Regarding SwarmUI Latest

  • https://youtu.be/FvpWy1x5etM

• Full FLUX Tutorial - FLUX Beats Midjourney for Real

  • https://youtu.be/bupRePUOA18

• Main Windows SwarmUI Tutorial (Watch To Learn How to Use)

  • How to install and use. You have to watch this to learn how to use SwarmUI

  • Has 70 chapters and manually fixed captions

  • https://youtu.be/HKX8_F1Er_w

• Cloud Tutorial (Massed Compute - RunPod - Kaggle)

  • If you don't have a powerful GPU or you want to use powerful GPU this is the tutorial you need

  • 48 GB A6000 GPU is only 31 cents per hour on Massed Compute with our special coupon

  • https://youtu.be/XFUZof6Skkw

• Free Kaggle Account Notebook for GPU-Poor

  • Installs latest version of SwarmUI on a free Kaggle account

  • Works with Dual T4 GPU at the same time

  • Supports SD 1.5, SDXL, SD3, FLUX, FLUX Tools, SD 3.5 Large, Mochi Tex to Video and Stable Cascade and more

  • Download from here : https://www.patreon.com/posts/106650931

18 February 2025 Update

• Hunyuan text-to-video support arrived to SwarmUI

• Public info link that has full information for how to manually install and execute from the SwarmUI developer below

  • https://github.com/mcmonkeyprojects/SwarmUI/blob/master/docs/Video%20Model%20Support.md#hunyuan-video

• SwarmUI Unified Downloader that we maintain (recommended)

  •  https://www.patreon.com/posts/114517862

  • Unified downloader has so many features including auto download GGUF models (20 steps) and Fast-GGUF (6 steps) models and configuration screenshots for Hunyuan

4 December 2024 Update

• Full tutorial for FLUX Tools published : https://youtu.be/hewDdVJEqOQ

  • Outpainting, Inpainting, Redux Style Transfer + Re-Imagine + Combine Multiple Images, Depth and Canny

28 November 2024 Update

• Tutorial : https://youtu.be/hewDdVJEqOQ

  • Updated out-painting best config : Link

  • To fix mask line you need to do second inpainting pass on mask line with regular FLUX dev model config : Link

• New FLUX Tools added to the SwarmUI

• I recommend using our updated unified downloader to auto download (Windows, RunPod, Massed Compute) : SwarmUI unified downloader (auto downloads you do nothing)

• Download flux1-redux-dev.safetensors into :

  • SwarmUI/Models/style_models

• Download flux1-fill-dev.safetensors , flux1-canny-dev.safetensors , flux1-depth-dev.safetensors into

  • SwarmUI/Models/diffusion_models

• Download flux1-canny-dev-lora.safetensors , flux1-depth-dev-lora.safetensors into

  • SwarmUI/Models/Lora

7 November 2024 Update

• Newest SOTA Text-to Video model genmo/mochi-1-preview now supported by SwarmUI

  • You can download FP8 scaled unified model from this repo

  • Put the downloaded model into \SwarmUI\Models\Stable-Diffusion

• Alternatively use our updated SwarmUI unified downloader

• Mochi 1 Tutorial Prompts : GitHub Gist

26 October 2024 Update

• Full tutorial video published and if you also upvote this reddit post I appreciate that

  • https://youtu.be/-zOKhoO9a5s

• Stable Diffusion 3.5 Large published

  • https://stability.ai/news/introducing-stable-diffusion-3-5

  • Update your SwarmUI to the latest - make sure updated worked

• For Stable Diffusion 3.5 Large I have compared 62 prompts and working Samplers + Schedulers

  • You can download all full grids as a zip : click here

  • For realism : DPM++ 2S Ancestral + SGM Uniform - This takes 2x time compared to Eular

  • For stylization like anime : Eular + Beta

• Put FP16 and FP8 FLUX models into : \SwarmUI\Models\diffusion_models

• Put FP16 and FP8 Stable Diffusion 3.5 Large into : \SwarmUI\Models\Stable-Diffusion

• Put FP8 or FP16 T5 XXL and Clip-L and Clip-G into : \SwarmUI\Models\clip

• For FLUX Dev best sampler is iPNDM + DDIM Uniform

• Ultra fast and robust automatic models downloader for Windows, RunPod and Massed Compute : click to download

  • https://www.patreon.com/posts/114517862

  • 

• For manually download click this link to go to downloads repo

  • T5 XXL Rename to Skip Download : t5xxl_enconly.safetensors

• Test results

• As FP8 version of T5 XXL it will download t5xxl_fp8_e4m3fn_scaled.safetensors which is supposed to be more closers to FP16 version

  • Here comparison (excellent quality) : FP8_vs_FP8_Scaled_vs_FP16_T5_XXL_With_FLUX.jpg

• FP8 version of Stable Diffusion Large 3.5 contains Text Encoders FP8 versions as well it is from ComfyUI

  • Here comparison i think FP8 scaled works better : FP8_vs_FP16_SD_3_5_Large.jpg

• We also added newest experimental scaled FP8 version of FLUX Dev

  • I think scaled FP8 FLUX DEV better than FP8 raw : FP8_vs_FP8_Scaled_vs_FP16_FLUX_DEV.jpg

• FLUX vs SD 3.5 Large Comparison: FLUX_ vs_SD35.jpg

18 September 2024 Update

• install-linux.sh file updated and now auto installs ComfyUI manager extension

• Just wait backends to fully load it takes some time

• You can check progress from Server > Logs > Debug menu

3 August 2024 Update

• SwarmUI now supports newest black-forest-labs FLUX models

• Full tutorial here : >> https://youtu.be/bupRePUOA18

• black-forest-labs has the original developers that developed first Stable Diffusion models

• I am recording a tutorial video

• You can find automatic downloaders for FLUX models at below post for both Windows, RunPod and Massed Compute

• 1-Click FLUX models downloader scripts for Windows, RunPod and Massed Compute are in below post including FP16 T5 Downloader file

• https://www.patreon.com/posts/109289967

• The above downloader scripts have resume capability, skip if fully downloaded capability and will auto retry 100 times if an error occurs and continue from where ever left automatically and downloads from Hugging Face directly

• Or you can instructions you can see at this image : How To Download And Use FLUX Models On Swarm UI.png

• FLUX models official repo : https://huggingface.co/black-forest-labs

• FLUX.1-schnell is timestep distilled - turbo model - 4 steps or more

• FLUX.1-dev is guidance distilled - best weights public model - 20 steps or more

• Non-distilled model is only available via API

• FLUX works with only CFG 1 - thus not supporting negative prompts at the moment

• FP8 Model works same as FP16-dev model if you are not manually using FP16 sampling (requires 28 GB) : https://huggingface.co/Kijai/flux-fp8/tree/main

• FP8 vs FP16 vs Turbo vs FP16 T5 model detailed comparison (16 different tests) : https://imgsli.com/MjgzODYy/7/6

• FP16 T5 Text Encoder can be downloaded from : https://huggingface.co/comfyanonymous/flux_text_encoders/tree/main

21 June 2024 Update

• Stable Swarm UI migrated into SwarmUI : https://github.com/mcmonkeyprojects/SwarmUI

• Links are updated

Update Error Fix

• fatal: detected dubious ownership in repository at 'G:/swarm_ui_tutorial/SwarmUI'

• If you get above error run below config (fix according to your install path)

• git config --global --add safe.directory G:/swarm_ui_tutorial/SwarmUI

General Info

• To download and install StableSwarmUI : StableSwarmUI

• Very important documents : docs

• Very important features : Features

• Full prompting syntax : Prompt%20Syntax.md

• Download very best new upscalers : released-sorted-by-new

• All upscalers as a single zip file : all_models

• Helaman upscalers with thumbnails : helaman

• SD3 Download model : stable-diffusion-3-sd3

• SD3 Official Page : stable-diffusion-3-medium

• SDXL model RealVis XL 4 : RealVisXL_V4.0.safetensors

• SD 1.5 model : Hyper Realism V3

• I have converted Fooocus 275 SDXL styles into StableSwarmUI presets with SD3 configuration

• The presets also have updated thumbnail / preview image with SD3

• Moreover, scripts to generate and update these presets automatically added to the zip file in below post as well

• You can download all presets from here : download 275 presets

• Official Stable Swarm UI Discord server to contact its developer : https://discord.gg/q2y38cqjNw

• Yolo Face segment into yolov8 folder : https://huggingface.co/Bingsu/adetailer/resolve/main/face_yolov9c.pt

• 

• Cloud tutorials are in production as well for Massed Compute, RunPod and hopefully a Kaggle

Massed Compute Instructions

• Tutorial : https://youtu.be/XFUZof6Skkw

• Please register Massed Compute from here : https://vm.massedcompute.com/signup?linkId=lp_034338&sourceId=secourses&tenantId=massed-compute

• After registering and loading balance go to this link : https://vm.massedcompute.com/deploy

• Select RTX A6000, if not available RTX A6000 [ALT Config] (same VRAM, GPU, Disk etc only RAM is 24 GB instead of 48)

• Then make the selections as below and deploy your machine

• This Massed Compute virtual machine image we have as SECourses has pre-installed SwarmUI with very best SD 1.5, SDXL and SD3 models

• You can also increase your GPU quantity to utilize all of them at once with SwarmUI

• Don't forget to run SwarmUI (1-click button) update before start using it

• If our image gets broken reply to this post and let me know - we will update it ASAP

• Also don't forget to apply Coupon code as SECourses to get A6000 GPU as 31 cents per hour (on RunPod cheapest 69 cents per hour)

How To Connect SwarmUI from your PC that is Running on Massed Compute - Preffered Cloudflare Way

First open a terminal and execute below commands to install cloudflared

• wget https://github.com/cloudflare/cloudflared/releases/download/2024.8.2/cloudflared-linux-amd64.deb

• sudo dpkg -i cloudflared-linux-amd64.deb

Then run SwarmUI to update to update it latest, then close the started SwarmUI terminal

Then open a new terminal and execute below commands and it will start SwarmUI and will give you a cloudflare link like (my-pills-sailing-pad-netherlands.trycloudflare .com) that you can connect

• cd /home/Ubuntu/apps/StableSwarmUI/

• ./launch-linux.sh --launch_mode none --cloudflared-path cloudflared

How To Connect SwarmUI from your PC that is Running on Massed Compute - I dont recommend 1 GB Bandwith

• Get your token from here for free : https://dashboard.ngrok.com/get-started/your-authtoken

  • sudo snap install ngrok

  • ngrok authtoken replace_with_your_ngrok_token

  • ngrok http 7801

RunPod Instructions

Follow this tutorial step by step to install on RunPod machine : https://youtu.be/XFUZof6Skkw

Please use this link to register : https://runpod.io?ref=1aka98lq - (it gives me credits thus i can do more tutorials)

• RunPod massive tutorial includes runpodctl for fast download to PC : https://youtu.be/QN1vdGhjcRc

• RunPod permanent network storage system tutorial : https://youtu.be/8Qf4x3-DFf4

• How to download models very fast to RunPod and how to upload files, folders from RunPod to Hugging Face ultra fast detailed tutorial : https://youtu.be/X5WVZ0NMaTg

Choose below template and any GPU

• RunPod Pytorch 2.2.0

• runpod/pytorch:2.2.0-py3.10-cuda12.1.1-devel-ubuntu22.04

Download install-linux.sh and upload into workspace

• chmod +x install-linux.sh

• ./install-linux.sh

It will auto install and start the SwarmUI

Then connect from 7801 port

• If you want to fix No module named 'onnxruntime' error which is 100% caused by SwarmUI itself while the app is running execute below commands after uploading install-linux_fix_step2.sh file

  • chmod +x install-linux_fix_step2.sh

  • ./install-linux_fix_step2.sh

• After executing and installing missing libraries, disable your back-end and enable again for it to become working

If you turn off your pod and start again execute below command to start again

• cd SwarmUI

• cd launchtools

• wget https://dot.net/v1/dotnet-install.sh -O dotnet-install.sh

• chmod +x dotnet-install.sh

• ./dotnet-install.sh --channel 8.0 --runtime aspnetcore

• ./dotnet-install.sh --channel 8.0

• cd ..

• ./launch-linux.sh --launch_mode none --host 0.0.0.0

SwarmUI Windows Requirements

• https://dotnet.microsoft.com/en-us/download/dotnet/8.0

• Git : https://git-scm.com/downloads

• How to install Git, Python, CUDA, cuDNN, C++ Tools : https://youtu.be/DrhUHnYfwC0

Error and Solutions

• If you get below errors

  • RuntimeError: CUDA error: operation not supported

  • CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.

• Add this parameter to the back-end and try : --disable-cuda-malloc

RunPod Cloudflared

• wget https://github.com/cloudflare/cloudflared/releases/download/2024.8.2/cloudflared-linux-amd64.deb

• dpkg -i cloudflared-linux-amd64.deb

• git clone https://github.com/mcmonkeyprojects/SwarmUI

• cd SwarmUI

• cd launchtools

• wget https://dot.net/v1/dotnet-install.sh -O dotnet-install.sh

• chmod +x dotnet-install.sh

• ./dotnet-install.sh --channel 8.0 --runtime aspnetcore

• ./dotnet-install.sh --channel 8.0

• cd ..

• ./launch-linux.sh --launch_mode none --cloudflared-path cloudflared